Dung Detective

Saving hundreds of thousands of animals every year by snapping a quick dung photo for instant health risk alerts

Comment

Inspiration

Pet owners often miss early warning signs of serious health conditions in their animals' waste. We realized computer vision could provide instant health screening that could save lives and reduce emergency vet visits.

What it does

PetHealth AI analyzes photos of pet waste to detect health issues across 9 animal species. It provides risk assessment, detailed health analysis, and emergency guidance using either offline image processing or cloud AI services.

How we built it

React/TypeScript frontend with Node.js backend. We implemented pure JavaScript computer vision (color histograms, edge detection, texture analysis) as the primary method, with optional integrations to 7 AI providers including Google Cloud Vision and GPT-5.

Challenges we ran into

The biggest challenge was avoiding API dependency - most AI solutions require expensive calls and internet. We solved this by building offline image processing algorithms. Getting species-specific health indicators right for 9 different animals was also complex.

Accomplishments that we're proud of

Creating a fully functional offline image analyzer that detects real medical indicators like blood, parasites, and dehydration through pixel analysis. No API keys or internet required for the core functionality.

What we learned

Computer vision doesn't always need neural networks - algorithmic approaches can be highly effective for specific medical use cases. Pure JavaScript image processing is surprisingly powerful for pattern recognition.

What's next for Dung Detective

Partner with veterinarians to validate detection algorithms, add historical health tracking, and expand to livestock farm management systems for large-scale animal health monitoring.

Built With

Share this project:

Updates